Half of adults worry as AI spreads; security firms urge use

A recent Pew Research survey found roughly half of adults say the increased use of artificial intelligence in daily life makes them more concerned than excited. Concern has grown over the past few years, driven by worries about jobs, reduced creativity, strained relationships, misinformation and more surveillance.

Survey participants reported higher awareness of visible technologies such as driverless cars and facial recognition, and lower awareness of algorithmic uses in welfare assessments, loan decisions and care services. Public concern about many AI applications has risen since 2022.

Security firms and researchers report that both defenders and attackers are adopting AI. Defensive tools apply machine learning to sift through large volumes of logs, detect unusual behavior, prioritize alerts and identify threats faster than manual methods. Attackers use AI to craft more convincing phishing messages, clone voices, generate realistic fake images and videos, automate reconnaissance and develop malware designed to evade older detection systems. Both sides use AI‑assisted methods to search for software vulnerabilities.

A 2025 study on AI in security found the public recognizes technical benefits in AI‑driven defenses-speed, scale and improved accuracy-while also expressing concerns about privacy, bias and potential job losses in security operations. When asked where they would invest unlimited money in AI research, respondents most often chose medicine and cybersecurity.

Surveys by security companies covering the US, UK, Austria, Germany and Switzerland show rising uncertainty about what to trust online and growing concern about impersonation and AI‑generated deception. Few respondents said they would be comfortable if defensive teams avoided AI while attackers continued to use it, because that would leave defenses slower and less effective.

Security professionals advise strict data hygiene when using AI tools. They recommend not pasting passwords, customer data or sensitive incident details into public AI services, treating AI outputs as unverified until checked against logs or source records, keeping a human decision‑maker in the loop for actions that affect access or legal responsibilities, and preferring enterprise or local deployments that offer logging, access control and clear data‑retention rules. They also warn that automated systems can produce confident but incorrect results and that every AI output used in security work should be validated against primary evidence before action is taken.